Quantum computing is usually explained with physics: superposition, entanglement, qubits. When I started reading about how superconducting quantum computers are actually built, what struck me was how much of the challenge is materials science and manufacturing. Coherence times, the time a fragile quantum state survives, are often limited by materials: impurities, surface oxides, interfaces, trapped magnetic flux, and losses in the dielectric layers and surfaces of the device.
As someone who has spent years on metal additive manufacturing, I wanted to know where AM could help, and where it would be the wrong tool. This article is my current view. It is a perspective from manufacturing, not from quantum physics.
The hardware in a few sentences
Superconducting quantum processors operate at millikelvin temperatures inside dilution refrigerators. Qubits are usually patterned circuits on chips, made by thin-film deposition and lithography, much like semiconductor devices. Around them is a large amount of hardware: packaging, shielding, microwave lines, filters, thermal links, mounts and, in some architectures, three-dimensional microwave cavities machined from bulk superconducting metal.
Those 3D cavities matter because a well-made cavity can store microwave photons for a long time. Some approaches use the cavity itself to hold quantum information, protected by error-correcting codes encoded in its states. In those designs, the quality of the cavity is central to the quality of the computer.
Why materials and manufacturing matter
For a superconducting cavity, losses come from a few places:
- The bulk metal, through impurities and defects that affect superconducting performance.
- The surface, where oxides, contamination, roughness and damage from machining add loss. Microwave currents flow in a very thin surface layer, so the surface dominates.
- Seams and joints, where two pieces meet and currents must cross an imperfect contact.
- Magnetic flux trapped as the metal cools through its superconducting transition.
Every one of these is a manufacturing question: material purity, forming and machining, chemical etching and polishing, heat treatment, joining, and cleanliness.
Where AM might fit
Additive manufacturing offers geometric freedom: internal cavities without seams, complex coupling geometries, integrated cooling and mounting features, and shapes that are difficult or impossible to machine. For quantum hardware, a few areas look promising to me.
Seamless or reduced-seam cavities. Many cavities are machined from two halves and joined, and the seam can add loss. A printed cavity could, in principle, avoid some seams or move them to places where currents are low.
Packaging and shielding. Sample holders, enclosures, filters and mounting brackets for cryogenic systems often have complex shapes and modest electromagnetic requirements. These are reasonable places to evaluate printed metal parts, provided they are clean, non-magnetic where needed and thermally adequate.
Thermal hardware. Cryogenic heat exchangers and thermal links benefit from the internal channels and high surface areas AM is good at. Copper and its alloys, which AM can now process, are natural candidates.
Rapid prototyping of geometries. Even if the final part is machined, printing test geometries quickly could speed up design iteration.
Where it clearly does not fit yet
I want to be equally clear about the limits.
- As-built surfaces are rough. LPBF surfaces are many times rougher than a well-machined and chemically polished surface. For a high-performance cavity, extensive surface processing would be needed, and the internal surfaces that make AM attractive are the hardest to finish.
- Purity is a challenge. The best cavities use very pure metals. Powder carries oxygen and contamination, and the printing atmosphere can add more. Achieving bulk purity comparable to high-grade wrought metal is not routine.
- Porosity and defects that are harmless in a structural part may matter in a superconductor, especially near the surface.
- Qualification data is thin. Printed superconducting resonators and cavities have been explored in research, but the evidence base is small compared with decades of experience on machined and formed parts.
So I would not claim that AM is an established route for high-coherence quantum components. I would say it is a promising research direction for some parts, and a practical option now for some of the surrounding hardware.
What a sensible AM program would look like
If I were starting such a program, I would:
- Start with the surrounding hardware: mounts, thermal links, enclosures, where requirements are clearer and wins are achievable.
- Treat surface processing as part of the process. Plan chemical polishing, heat treatment and cleaning from the first design, and design geometry so internal surfaces can be reached.
- Measure what matters. Superconducting performance at operating temperature, not only density and tensile strength.
- Work with the physicists early. The requirements for a cavity are set by microwave and quantum performance, which manufacturing engineers do not measure every day.
- Compare honestly against the best machined parts, not against an average one.
Quantum hardware is one of the most demanding applications imaginable for manufacturing. That is exactly why it is interesting. The companies that build these machines will need people who understand both the physics and the factory floor. In the next article I look more closely at the two metals that dominate superconducting cavities: aluminium and niobium.
Further reading
- Reagor, M. et al. (2013). Reaching 10 ms single photon lifetimes for superconducting aluminum cavities. Applied Physics Letters 102, 192604.
- Romanenko, A. et al. (2020). Three-dimensional superconducting resonators at T < 20 mK with photon lifetimes up to τ = 2 s. Physical Review Applied 13, 034032.
